Fuzhou tone sandhi

My PhD focus is on Fuzhou, a Min Chinese variety spoken in the city of Fuzhou, Fujian, China. In this variety, the right-dominant tone sandhi determines the pre-final sandhi tones based on the identity of both the underlying tone and the rightmost context tone. Massive neutralisation in the tonal paradigm makes it interesting to look at just how categorical the neutralisation is acoustically, and how productive these "phonological rules" are in wug-tests. I am also creating a socially stratified corpus of Fuzhou Min speech, in a sociotonetic study of the interaction among tone sandhi, tonal coarticulation, intonation and focus.

A key part of my PhD looks at incomplete neutralisation in Fuzhou tone sandhi. The phenomenon occurs when two tones which are contextually neutralised to sound like the same tone, turn out to be acoustically different in sutble ways. Using Bayesian hierarchical modelling, whose assumptions allow us to both reject and confirm the null hypothesis, I show that tonal neutralisations in Fuzhou are sometimes complete, sometimes incomplete. This raises questions about the amount of phonetic details that must be stored in phonological representation. My preferred solution is a phonetically-rich representation (exemplars, if you like) which nevertheless allows categorical effects to emerge.

Cross-linguistic prominence perception

I am leading a project on the role of acoustic cues when listeners of different languages perceive prosodic prominence in speech. To enable a cross-linguistic comparison, it is important to deploy the same sets of stimuli and procedure for listeners with varied linguistic backgrounds. We test 8 typologically distinct and related varieties: French, Cantonese, Fuzhou Min, Singaporean English, Standard Southern British English, Danish, and Valais and Schaffhausen Swiss German. 170 listeners of these varieties are tested on 5 groups of stimuli, consisting of disyllabic [baba] sequences where the duration, amplitude envelope and pitch contour of both syllables are systematically manipulated. Listeners are asked to identify which of the two syllables they hear as “stronger”. We can then see the overall bias in each variety towards selecting the first or the second syllable and the sensitivity to changes in acoustic cues for each variety.

Dr Adrian Leemann (Lancaster)
Dr Marie-José Kolly (Zurich)
Dr Ricky Chan (Lancaster)
Dr Anna Jespersen (Aarhus)
Geraldine Kwek (Cambridge/NIE Singapore)

Variation and change in Mandarin dialects

My interests in Chinese dialectology and sociolinguistics have led to projects studying variation and change among contemporary Mandarin varieties, in Dalian, Beijing and beyond, as dialect speakers respond to rapid urbanisation and pressures of standardisation. One of the things we discover is that convergence to the standard is by no means the only destination for these dialects, particularly when local identities can be constructed around local features. More work is planned on the sociophonetics of and changes across the lifespan in Mandarin dialects.

Shichao Wang (Dalian Tech)
Chris Xia (Cambridge)
Arthur Thompson (Hong Kong U)

Cantonese tonal coarticulation

Dr Ricky Chan (Lancaster) and I are working on tonal alignment and tonal coarticulation in Hong Kong Cantonese, testing the effects of syllable composition and speech rate. We are currently analysing data and incrementally presenting our results.

English Dialects App

Since my BA in English Language I have been interested in the variation and history of English. Recently I've had the opportunity to analyse data from the English Dialects App, created by Dr Adrian Leemann (Lancaster), Prof David Britain (Bern) and the team. I've worked with Tam Blaxter (Cambridge) and Benjamin Gittelson in this effort, producing population statistics for acoustic laryngeal voice quality in the App corpus.